Loading...
Back to Portfolio
Mobile DevelopmentClient Project
Smart Home Control App
One Flutter app unifying a smart-home device line: MQTT control, learned routines, voice assistants
The client's name is withheld. Ask us to walk you through this build live.
Client
A smart-home device manufacturer
Timeline
5-month build
FlutterFirebaseMQTTAWS IoTTensorFlow Lite
About This Project
A smart-home device manufacturer had a separate app for each product category: lights, thermostats, cameras, and sensors. This Flutter app development project consolidated the whole line into one cross-platform application. Devices communicate over MQTT topics brokered through AWS IoT Core, and scenes orchestrate multiple devices as a unit. TensorFlow Lite models running on the phone learn household routines and propose automations. Alexa, Google Assistant, and Siri Shortcuts are integrated natively for hands-free control.
The Engineering Problem
Unifying four separately developed device lines is as much a protocol problem as an app problem. Each category spoke its own dialect, and pairing flows failed on congested home networks. Automation had to work across device boundaries the individual firmwares never anticipated.
- •Four device categories with divergent protocols and account systems to merge into one model
- •Device pairing must survive congested 2.4GHz Wi-Fi and flaky home networks
- •Cross-device scenes need an orchestration layer no single firmware provides
- •Routine learning must run on the phone; streaming raw home telemetry to the cloud was off the table
- •Each voice assistant imposes its own device and intent model on the integration
Technical Approach
We unified device communication behind an MQTT topic architecture and rebuilt pairing as a guided, fault-tolerant flow. The intelligence stays on the device: routine models run locally instead of on raw telemetry shipped to the cloud.
- Designed a unified MQTT topic scheme with per-device shadows through AWS IoT Core
- Rebuilt pairing with BLE-assisted setup, retry logic, and clear failure recovery at every step
- Built a scene engine orchestrating multi-device actions with ordering and rollback
- Embedded TensorFlow Lite routine models that run fully on-device with federated updates
- Integrated Alexa, Google Assistant, and Siri Shortcuts against one shared capability model
How the Build Ran
Device Protocol Unification
2 weeksMapped the protocols across all four device lines and designed the MQTT topic and device-shadow architecture.
App Development
12 weeksBuilt the Flutter application (pairing, real-time control, scenes, and the automation engine) with continuous testing on both platforms.
On-Device ML
4 weeksTrained and embedded the routine-learning models with TensorFlow Lite and wired suggestion flows into the app.
Beta & Hardening
2 weeksRan a structured beta across hundreds of households, refined the setup flow, and tuned battery and reconnection behavior.
Key Features in This Build
Unified Device Control
Every device category in one app with room-based organization and live state
Scene Orchestration
Multi-device scenes with ordered execution and safe rollback
On-Device Routines
TensorFlow Lite models learn household patterns locally and suggest automations
Voice Assistants
Alexa, Google Assistant, and Siri Shortcuts against one capability model
Build Scope
We designed and delivered the scope below on this engagement, and we can walk you through the app, the protocol layer, and the integrations live.
4
Device categories
Lights, climate, cameras, and sensors unified
3
Voice assistants
Alexa, Google Assistant, and Siri Shortcuts
MQTT
Transport
Device shadows brokered through AWS IoT Core
2
Platforms
iOS and Android from one Flutter codebase
Technology Stack
mobile
- Flutter
- Dart
- Provider
- Flutter Bloc
backend
- AWS IoT Core
- Lambda
- DynamoDB
- Firebase Cloud Messaging
iot
- MQTT
- Device Shadows
- Zigbee Bridge
- BLE
on-device-ml
- TensorFlow Lite
- Firebase ML
- Edge Inference
- Federated Learning
Designed to Extend
Extensions this architecture is built to support:
- Matter protocol support for cross-ecosystem device compatibility
- Energy management modules with utility API integration
- Tablet and smart-display companion apps on the same codebase
- Predictive maintenance alerts built on the routine-learning models
Want to See It Running?
Ask us to walk you through this build live, and to map the approach onto your product.
Start Your Project